Domhoff, G. William (1996). Finding Meaning In Dreams: A Quantitative Approach. New York: Plenum.
 • Introduction
 • Chapter 1: The Scientific Study of Dream Content
 • Chapter 2: The Hall/Van De Castle System
 • Chapter 3: The Quality of the Data
 • Chapter 4: Normative Findings
 • Chapter 5: Age Differences in Dream Reports
 • Chapter 6: Cross-Cultural Studies of Dream Content
 • Chapter 7: Consistency and Change in Long Dream Series
 • Chapter 8: The Continuity Between Dreams and Waking Life
 • Chapter 9: The Repetition Dimension
 • References

Chapter 5: Age Differences in Dream Reports


Age is a defining issue in the waking lives of all Americans. It is now time to see if this is also the case in their dreaming lives. Reversing the usual order for discussions of age, we will start with adults and end with children. The chapter begins by comparing college-age findings with those for post-college groups. It next turns to teenagers (defined as ages 13-17) and then to children (defined as ages 2-12). The longest section of the chapter concerns a detailed comparison of conflicting findings on the dream reports of young children. This discussion shows how important the adoption of a common coding system could be in avoiding misunderstandings, but it also shows that with children there are likely to be larger differences between dreams collected in the home and laboratory than we think is the case with adults.

Young Adults Versus Older Adults

Using an earlier version of the Hall/Van de Castle coding system that varies only slightly from its final form, we did studies in the early 1960s of characters, aggressions, and friendly acts in dream reports of young and older adults. These studies allow us to make comparisons between college students, all of whom were between 18 and 25, and older adults who ranged in age from 30 to 80 (Hall and Domhoff, 1963a, 1963b, 1964). The older adult sample was developed by taking one dream from dream series provided by 281 males and 281 females. These dream series were given to Hall by a wide variety of people.

The main age-related findings in these studies are brought together in Table 5.1. There is stability in the male/female percent, but a decline in friendly interactions and aggressions.

Table 5.1. Age Differences in Selected Content Categories for Young vs. Old Adultsa
       Males    Females  
Male/Female percent 66/34 (68/32)  52/48 (55/45) 
F/C index overall13 (21)04 (14)
A/C index overall22 (34)07 (24)
A/C index with male characters18 (27)03 (17)
A/C index with female characters05 (14)03 (13)
a There were 16 possible combinations for settings.
h = .21 between younger men and over-30 men on F/C overall.
h = .34 between younger women and over-30 women on F/C overall.
h = .49 between younger women and over-30 women on A/C overall.

Harold Zepelin (1980-81, 1981) did a very good and thorough study of age differences in 58 males ages 27 to 64 using both laboratory and home dream reports. He used both his own scales and some of the Hall/Van de Castle scales. There were very few age differences. As in the Hall and Domhoff (1963b) study, there was a slight age-related decline in aggression. There also was decline in "distortion" on a scale designed by Zepelin. Family-related content was most prominent from ages 35 to 55, which makes sense in terms of the focus on child rearing for most adults in those years. However, Zepelin stresses how few changes there were in the several dozen comparisons he made, and notes that the correlations were low even where there were changes.

As one part of an in-home mental health survey conducted in Cincinnati with a representative sample of 300 men and women age 21 or over, the interviewers asked the subjects to relate their most recent dream (Kramer, Winget, and Whitman, 1971; Winget, Kramer, and Whitman, 1972). Sixty-four men (53% of the male sample) and 118 women (65% of the female sample) provided a total of 182 dream reports. The dream reports were compared for gender and age differences on the major Hall/Van de Castle scales. Four age classifications were used: 21-34, 35-49, 50-64, and 65 and over.

Most of the usual gender differences were found. The women's reports contained more characters, friendly interactions, emotions, indoor settings, and mentions of family, while the men's dream reports had more aggression and successful striving (Winget, Kramer, and Whitman, 1972:203). However, the only age difference was a lower amount of aggression for those in the 35-49 age category as compared to those who were younger or older. There is one reason for caution with these findings. The mean length of the dream reports was only 21 words, with a range of two to 68 words. They were thus far shorter than the reports used by Hall and Van de Castle to establish their norms; their dream reports averaged about 125 words, with a range of 50 to 300 words. After noting the differences in the length of reports, Kramer, Winget, and Whitman (1971:89) conclude "the striking feature that emerges is how similar the dream content of the two samples turns out to be."

Brenneis (l975) compared 148 dream reports from 38 college women ages 18 to 26 (mean age: 20.5) with 185 reports from 43 women ages 40 to 86 (mean age: 56.8). The dreams of the older women were obtained from friends of the author and friends of friends (a "snowball" sample). The dream reports of the younger women were obtained through college classes. All but one of the older women had attended college. All but one was or had been married; eight were widows. Brenneis compared the two sets of dream reports on his ego dimension scale briefly described in chapter 2. However, most of his comparisons are subjective and interpretative. For our purposes here, his main finding was that there was less aggression in the older women's reports, with levels of robust and energetic activities remaining high. There was no decline in sexual elements (Brenneis, 1975:433).

A study of how college-age and elderly women (59-87 years) rated the emotionality of their dream reports supports the idea that there is a decline in aggressive/hostile dream content in older adults (Howe and Blick, l984). Both sets of women kept dream diaries for six weeks and rated each dream on a ten-item emotional checklist developed for dream studies by Stairs and Blick (1979). The authors found anger-rage and fear-terror were checked significantly less often in relation to the dream reports of the elderly; enjoyment-joy accounted for a significantly higher proportion of emotions related to the dream reports of the elderly.

The findings with American subjects are supported in a study of 47 French-Canadian women who ranged in age from 25 to 56, each of whom contributed two dream reports to the study (Lortie-Lussier, 1995). Seventeen of the women were from 25 to 35, 20 from 36 to 45, and 10 from 46 to 56. The dream reports were scored on the Hall/Van de Castle scales for characters, settings, activities, emotions, and aggressive and friendly interactions, and then sorted into the three age groups. There was a slight decline in the total number of emotions, and more pleasant outcomes in the oldest group, but no other age differences. Although we cannot be confident of these findings with only 20 to 40 dream reports in each age group, they point in the same direction as the studies of older Americans.

Ideally, these cross-sectional findings should be checked against longitudinal studies to make sure the few differences are not due to cohort effects. But, needless to say, systematic longitudinal studies of dream content over a period of years or decades are not likely to be undertaken. What we do have are a few multi-year dream series discussed in chapter 7 on consistencies in long dream series. To anticipate, the findings from those dream series support the cross-sectional findings in showing little or no change over the span of decades. Unfortunately, these few series are not helpful on the issue of changes in aggressive content because they come from mild-mannered people who started their dream journals in their mid-20s or early 40s with very low levels of aggression that remained constant thereafter.

Teenagers (13 to 17)

There is very little information on the dream reports of teenagers, and some of what we have is disappointing because it is not understandable in terms of the Hall/Van de Castle norms. Moreover, some of the trends, such as on aggressive interactions, are contradictory. There are also some reassuring consistencies, such as on characters, settings, and objects, but we need better studies before we can be very confident about extending the norms to teenagers. Since it is not easy to obtain dream reports on a regular basis from teenagers over the space of several weeks, studies relying on the Most Recent Dream approach outlined in the previous chapter may be the best way to conduct future studies of this age group.

Hall and Domhoff (1963b, 1964) analyzed one dream report from each of 138 male and 138 female teenagers, ages 13-18, as part of their larger study of aggressions and friendliness in different age groups. For our purposes here, we will compare those findings with their study of one dream report from each of 200 males and 200 females ages 18 to 27 rather than the Hall/Van de Castle norms because of small changes in the coding system made when the norms were created. The point for now is only to see how closely the teenagers approximate the young adults.

As can be seen in Table 5.2, there are some differences. The teenage males express more friendliness than the young adult males (F/C index). Consequently, the teenage males have a lower aggression/friendliness percent. They are also more likely than young adult males to be victims of aggression, but a smaller percentage of the aggression in their dream reports is physical. For women, the differences are in aggressions. As young teenagers they have a higher A/C code, a higher rate of victimization, and a higher percentage of physical aggression. Consequently, the aggression/friendliness percent for teenagers is slightly higher than it is for young adult women.

Table 5.2. Differences between Teenagers and Young Adults in Selected Content Categoriesa
       Males    Females  
Male/Female percent 66/34 (68/32)  52/48 (55/45) 
F/C index overall13 (21)04 (14)
A/C index overall22 (34)07 (24)
A/C index with male characters18 (27)03 (17)
A/C index with female characters05 (14)03 (13)
a There were 16 possible combinations for settings.
h = .21 between younger men and over-30 men on F/C overall.
h = .34 between younger women and over-30 women on F/C overall.
h = .49 between younger women and over-30 women on A/C overall.

There is a very rigorous dissertation comparing dream reports from high school and college students in Ames, Iowa, using the Hall/Van de Castle coding categories (Howard, 1978). Unfortunately, it did not use the Hall/Van de Castle methods of analyzing the data, so the results are not comparable with the Hall/Van de Castle norms.* The author compared dream reports from 44 15-16 year-olds (17 males, 27 females) at Ames High School with those from 54 20-22 year-olds (24 male, 32 female) in upper-division psychology courses at Iowa State University. All subjects recorded their dreams at home for one week in a booklet consisting of six lined pages. The difficulty in obtaining the high school reports is seen in the fact that only 32 of the 63 original high school volunteers returned any dreams; the figure was 13 out of 23 for the second volunteer group. One student returned only one report of nine words, and had to be dropped. By way of contrast, 54 of the 56 college volunteers returned the dream booklet (Howard, 1978:54). There were 381 usable reports in all.

There is no information on how these dream reports were distributed by age and gender. However, Howard (1978:36, 56) does report that the small number of high school males in the sample was compounded by the fact they produced the fewest dream reports and used the fewest words to describe their dreams (76 words per dream report vs. 117 for college males; by comparison, high school girls wrote an average of 105 words, college females 120 words; Howard, 1978:56). The high school boys were the least clear in their descriptions, and they had the lowest scores for many content categories. In short, they do not sound like eager participants, and comparison of them with the other three groups must be made with caution.

All dream reports were randomized, then rated independently by two judges on the following variables: dream length, setting, characters, aggression, friendliness, sexual content, references to the body, food, clothing, weapons, emotions, references to the past, negative affect, unusual elements, and negative words (Howard, 1978:34). Reports of fewer than 15 or more than 175 words were excluded. Eight of 389 reports were eliminated because they contained less than 15 words. None was eliminated for being too long (Howard, 1978:54).

Howard made her analysis so she would have one score for each subject for each coding category. This controlled for dream series of different lengths. Here is how she did it:

To obtain one score per subject on each variable, an average score per dream or mode was established first, then an average or mode of all the dreams produced by the subject was determined. This procedure established the number of subjects rather than the number of dreams as the basis for analysis (Howard, 1978:51).

Howard used chi square, pearson r, and analysis of variance to analyze her findings. She found very few age differences. The two age groups were similar on settings, objects, characters, friendly and aggressive interactions, and unusual elements. The high school students were higher on the descriptive and evaluative scales, meaning they used more modifiers like size, speed, and intensity to describe people, objects, and events. College students had more male figures in their dream reports. They also expressed more negative emotions and used more negative words (Howard, 1978:63, 96, 102, 106, 111, 138).

On the other hand, there were several gender differences. College males expressed more aggression than females. High school and college males had more outdoor settings, whereas high school and college females had more indoor settings and more indoor and outdoor settings in the same dream report (Howard, 1978:61, 90). High school and college females scored higher than males in female characters, friends, friendly interactions, and references to food (Howard, 1978:71, 78, 93, 138).

If we assume Iowa State University students are similar to those at the universities discussed in the previous chapter, an assumption supported by the gender differences reported by Howard, then we can conclude tentatively that the Hall/Van de Castle norms might be of use for teenagers as young as age 15. However, we need further studies before we could be confident of that conclusion. Specifically, given our own findings with young teenagers, and the fact that Howard did not control for gender differences in report length or number of characters, we are uncertain about her findings on aggressive and friendly interactions. We are not confident about our own findings, but we are not confident about hers either.

A laboratory study of teenagers' dream reports was conducted by Foulkes (1982) as the final part of a longitudinal study of dreaming and dream reports in children from age 3 to 15. It will be reported on in the next section on children's dream reports. It did not use the Hall/Van de Castle system, but some of its results are roughly comparable.

Children's Dream Reports

As part of their developmental study of characters, aggressions, and friendliness, Hall and Domhoff (1963a, 1963b, 1964) analyzed 217 dream reports from 119 boys ages 2-12 and 274 reports from 133 girls in the same age group. Hall later increased the sample size to 600 reports in an unpublished study. Findings from both analyses will be reported in this section.

These dream reports were collected for Hall by college students, parents, and teachers in home, nursery school, and school settings. There were usually only one or two dream reports from any one child. Most of the dream reports came from children over the age of 6. It is likely, as we will see, that this sample is biased toward memorable or frightening dream reports, so the findings need to be treated with more caution than the teenage findings, at least for some categories. Moreover, the children's reports are shorter than those of teenagers or young adults.

An analysis of the characters in the children's reports first of all showed that the gender difference on male/female percent was present--69/31 for the boys, 50/50 for the girls. Since we have so few dream reports from children under age 5, we also determined the male/female percent in the narratives of a sample of children ages 3 to 6 produced when they were asked to tell a story (Pitcher and Prelinger, l963). The results are very similar to what we found in dream reports, as shown in Table 5.3. The biggest difference between child and adult dreams on the character scales is the much larger number of animals in children's dream reports, which manifests itself as a higher animal percent. This high animal percent also appears in young children's stories, as shown in Table 5.3. Table 5.4 presents the animal percent findings in a separate unpublished study by Hall. They are not as high as in children's stories. The animal percent declines with age. In a later, even more detailed study, Van de Castle (1983) reported similar findings using dream reports he collected with the aid of teachers as well as dream reports in Hall's collection.

Table 5.3. Male Percent and Animal Percent in Children's Stories
       No. of Stories    Mean Characters    Male/Female Percent    Animal Percent  

Table 5.4. Animal Percent in the Dream Reports of Children Age 2-12
       No. of Dreams    Mean Characters    Animal Percent  
2 - 61061001.722.073224
7 - 121962001.902.131414

A higher animal percent in children's dream reports also was found in the first laboratory study of children's dream reports. Foulkes et al. (1967) collected dream reports from 32 boys ages 6 to 12 who slept two nights in the laboratory and compared them to laboratory reports from young adults. Animals accounted for 16% of the characters in the boys' narratives; no animals were found in the reports of the young adult males.

Turning now to social interactions, Hall and Domhoff (1963b) found there is more aggression in children's dream reports than in adults' reports in terms of the A/C index. Much of this larger amount of aggression is with animals, and the child is usually the victim of an attack by the animal. After animals, males are the most frequently involved character in the aggressions in children's dream reports. The rate of aggressions that girls have with male characters declines into adulthood, but there is no difference between boys and college males in this respect. In terms of the victimization percent, children are more likely to be victims of aggression than are adults. In keeping with these findings on higher rates of aggression, children in this study also have more dreams with at least one misfortune than do adults. The F/C index shows there is almost as much friendliness in children's dream reports as in those of adults. For boys, the rate of friendliness with other males increases with age. For girls, the rate of friendliness with both genders stays the same as they grow older.

To summarize, the key differences between the children's sample and adult dream reports are as follows:

  • children have a higher animal percent;
  • children have a higher A/C score;
  • children have a higher victimization percent;
  • children have a higher percentage of dreams with at least one misfortune in them.

As stated at the outset of this section, these findings on children may be biased by a very selective sample. It is therefore necessary to see how they compare with findings from an extensive developmental study of dreaming in the sleep laboratory.

The Foulkes Study

One of the best studies of dreams and dreaming ever conducted in the sleep laboratory, and by far the most comprehensive study of dreaming in children, was carried out over a five-year period at the University of Wyoming by David Foulkes (1982). The basic study involved two groups of children who slept in the laboratory every other year for nine non-consecutive nights over a five-year period. The first group was between 3 and 4 years of age when it started. The second group was between 9 and 10. Thus, Foulkes has information on the dream reports of children ages 3 to 15. In addition to the nine regular nights in the laboratory, in which subjects were awakened three times per night, Foulkes gave the children a wide range of cognitive and personality tests, observed them at a summer play school, interviewed their parents, and had them write down their home-recalled dreams at one point in the study. The study began with 30 children in the two groups; six boys ages 11 to 12 were added at the start of the third year, and seven girls ages 7 to 8 were added at the start of the fifth year. In all, 46 children were studied--26 for all five years, 34 for at least three years, and 43 for at least one complete year. Foulkes made 2,711 awakenings over the five-year period.

Perhaps the most surprising finding of the study was the low amount of recall from REM periods in the 3 to 5 year olds, and the static, bland, and underdeveloped content of the few reports that were obtained. Only 27% of the REM awakenings yielded any recall that could reasonably be called a dream. The reports became more "dreamlike" (characters, themes, action) in the 5 to 7 year-olds, and increased in frequency in the 7 to 9 year-olds, but it was not until the children were 11 to 13 years old that their dreams began to resemble those of adult laboratory subjects in frequency, length, and overall structure (Foulkes, 1982:217).

Using a cross-sectional design, Foulkes later replicated his most important findings at Georgia Mental Health Institute with 80 children, ten boys and ten girls of ages 5, 6, 7, and 8. He and his co-authors summarized their findings as follows:

Dreams were reported relatively seldom (median report rate of 20%); until age 7, their imagery was reported as more static than dynamic; until age 8, a passive-observer role for their self character was most common; until age 8, dream activity evidenced very simple forms of narrative structure; waking visuospatial, but not verbal, skills predicted dream-report rates, with Wechsler Block Design the single best such predictor (Foulkes, Hollifield, Sullivan, Bradley, and Terry, 1990:447).

We now need to compare Foulkes's findings with those from the Hall collection, especially because children's dream reports do not seem to be as full of aggression and misfortune as the Hall and Domhoff findings would lead us to expect. However, it is difficult to make a detailed comparison because Foulkes utilized his own coding system, and it does not correspond completely with the Hall/Van de Castle system. We therefore have to piece together a comparison with great care.

First, we do know that the gender difference in male/female percent is present in Foulkes's dream reports because Hall coded them on this dimension (Hall, 1984:1115). Since the reports of the younger children were short and contained few characters, Hall combined the data for ages 3 to 9. The percentage of male characters in 209 male dream reports was 70/30 as compared to 51/49 in 146 female reports. In the longer and more numerous narratives of the 9 to 15 year olds, there was no systematic variation from year to year in the male/female percent, so Hall combined them into one group. The 510 male dream reports in this group showed an even higher male percent (76/24), whereas the 496 female reports had an even lower male percent (43/57).

Second, we know that animals appeared in a very high percentage of the young children's reports in the Foulkes study, and this percentage declined with age, as it did in the study by Van de Castle (1983). Foulkes (1982:48) reports that animals were the "major" characters in the dream reports of children 3 to 5. They appeared in 45% of girls' dream narratives and 33% of boys' reports. Animals appear in 38% of REM reports for 5 to 7 year olds, most of whom by then are having narrative-like dreams with social interaction and physical movement (Foulkes, 1982:80). At this age period boys are dreaming more often of animals than are girls. In the years 7 to 9, a decline in dreams with animal figures in them begins (Foulkes, 1982:113, 115, 234).

Third, there was at least some aggression in the dream reports in the Foulkes study. Between the ages of 5 and 7, one dream in four for both boys and girls "contained a hostile (attack) initiation by one character to another"(Foulkes, 1982:9-90). At this age all but one of these hostile attacks was initiated by some character other than the dreamer and the dreamer was rarely the victim. The same low level of aggression was found at ages 7 to 9, although some of the acts were now initiated by the dreamer--in 9% of the girls' dream reports, 4% of boys' reports (Foulkes, 1982:120).

Within this context of low overall levels of hostility, there were nonetheless gender differences. At ages 5 to 7, girls had far more reports with nonhostile social interactions (56% vs. 21%), and at ages 11 to 13 the "prosocial initiations" were higher in girls than boys (Foulkes, 1982:89, 189-90). Conversely, boys 11 to 13 had more "antisocial acts or misfortunes" (Foulkes, 1982:189-90). The increased gender difference at ages 11 to 13 was due to a decline in aggressions in the dream reports of the girls (Foulkes, 1982:193).

In the years between 13 and 15, the dream reports of Foulkes's subjects became more like the Hall/Van de Castle norms. First, there was less prosocial behavior and more ascriptions of anger to other dream characters (Foulkes, 1982:229). For girls there was an increased amount of victimization (Foulkes, 1982:231). There was a general decline in dreaming about family members, especially for males, and the tendency for girls to be in home or residential settings became statistically significant (Foulkes, 1982:230, 235, 247).

Overall, the findings on the male/female percent and the animal percent, and the gender differences on family members, aggressions, and settings, are similar to what was found by Hall and by Hall and Domhoff. However, there appears to be less aggression, misfortune, and negativity in the Foulkes data. Between the ages of 11 and 13, for example, "prosocial initiations" outweighed "antisocial" ones by a ratio of two to one, and even in the less prosocial reports of 13- to 15-year-olds, there were more prosocial motives than antisocial ones (Foulkes, 1982:189, 231). By way of contrast, in the Hall and Domhoff (1963b, 1964) findings, there was more aggression than friendliness, although it should be noted that the F/C index was higher for teenage males than for either younger boys or male college students. Part of this difference between the two sets of findings may be due to differences in the coding systems, but it probably means Hall and Domhoff had selective samples of dream reports for children and teenagers.

In order to bring some new findings to bear on these differences, we asked Tonay to apply several Hall/Van de Castle coding categories to the dream reports Foulkes et al. (l990) collected in their cross-sectional replication study. Tonay coded the reports without knowledge of the gender or age of the dreamer. Table 5.5 shows the findings for settings. Both boys and girls tend to be outdoors or in no setting at age 5, but by age 7 the girls tend to be indoors and the boys to be outdoors. This finding would come as no surprise to either Hall or Foulkes. The number of dreams at each age level is relatively small, so we do not want to make any statistical claims.

Table 5.5. Percentage of Indoor, Outdoor, No Setting and Familiar Settings
       # of Dreams    Indoors    Outdoors    No Setting    Familiar  

Table 5.6 presents the main findings for characters. Consistent with the adult norms, there are more characters in the girls' dream reports. The usual findings on the male/female percent appear from age 5, and on the familiarity percent at age 7. The animal percents are not greatly different from those in Table 5.4.

Table 5.6. Findings with Characters
       # of Dreams    Mean number  
  of characters  

The findings on social interactions, misfortunes, and negative emotions are in basic agreement with Foulkes's earlier findings. Furthermore, Tonay found no failures, very few misfortunes, and few negative emotions. There is more friendliness in the girls' reports, but relatively little aggression in the boys' reports. Several of these findings are presented in Table 5.7. The number of aggressive and friendly interactions is too small to make statistical claims.

Table 5.7. Aggression and Friendliness
       A/C index    # of aggressions    # of friendliness    F/C Index  

When all the comparisons of the Hall and Foulkes data on children and young teenagers are taken into consideration, it is clear that Foulkes is right when he says our findings on aggression and misfortune in the dream life of children between the ages of 2 and 12 are based on a selective sample. At the same time, several of the adult gender differences appear in Foulkes's University of Wyoming and Georgia Mental Health Institute data by age 7 or 8, and other previously reported adult gender differences appear by the ages 13 to 15. We also note that he found the same developmental trends on the decline of animal percent, the increase in male strangers, and the victimization percent.

Thus, not all the differences are due to the nature of the samples. Some of them are due to differences in the coding systems. The Hall/Van de Castle coding system obtains somewhat different results because it has different coding rules for friendliness and because it corrects for the larger number of characters in girls' and women's dream reports.

Findings similar to those in the Foulkes data are being produced in an ongoing longitudinal study in Switzerland of 12 boys and 12 girls between the ages of nine and 15 (Strauch, 1995). For example, only a small percentage of their dream reports (12% to 24% when they were between ages 9 and 12) contained either a friendliness or an aggressive interaction. Nonetheless, dream reports with at least one aggression outnumbered those with at least one friendliness for both boys and girls by age 12, and the victimization percent began to approach that of young adults by the same age in the relatively few dream reports with aggressions. Once laboratory dream reports are collected from the subjects when they are ages 15 to 16, the final report from the study will include a comparison with their home-recalled dreams over the same six-year age span. The overall study promises to add significantly to our understanding of how dream content develops in the teenage years.

There are not likely to be many more large-scale studies of children's dreams in the laboratory because laboratory studies are expensive and time-consuming. The next feasible step in trying to resolve any remaining differences between the earlier studies therefore might be to see if the Most Recent Dream approach could be extended to younger children, perhaps with teachers reading them the instructions and asking them to write out their dream reports anonymously for the investigator while they are sitting at their desks. The younger the children, the more likely they might be to respond in terms of a memorable, recurrent, or frightening dreaming. However, even large samples revealing consistent findings at ages nine to 12 would be a useful addition to our understanding, especially if they could be coupled with Most Recent Dream studies of teenagers, where the method is definitely feasible.

Some indication of what might be possible is seen in a pilot study of Most Recent Dreams from 64 male and 80 female seventh graders, ages 12 to 13, in two middle schools in California (McNicholas and Avila-White, 1995). In keeping with adult differences, the reports from the girls were longer. Parallel to other findings mentioned earlier in the chapter, the animal percent was .13 for boys and .02 for girls, and the male/female percents were very close to the adult norms. The percentage of dreams with at least one aggression was 47 for the boys, which is exactly what it is for young adult males, and the figure for girls was 37, not far from the young adult female norm of 43. The percentage of dreams with at least one friendliness was 56 for the girls, not too far above the young adult female norm of 42. The only atypical finding was the low percentage of boys' dreams with at least one friendliness, 17, as compared with the normative figure of 38. These findings need to be replicated with larger sample sizes in a variety of settings, but they are an encouraging starting point.


We conclude that age is not a major factor in shaping dream content once Americans have reached young adulthood. One exception is the possible decline of aggressive acts in the dream reports of people over age 30. Thus, in studying individual dream series or dream sets from people over age 30, we should not be surprised by aggression scores lower than the norms. Instead, perhaps we should pay closer attention to aggression scores only slightly above the norms. In all, however, the important point is that norms based on college students are useful in studying post-college adults.

Still, the norms have to be used with a certain amount of common sense. For example, it should not come as a surprise if parents dream of their children, thus increasing the percentage of youthful characters over the young adult norms. Such findings are not criticisms of the norms, but evidence of their usefulness in developing our understanding of the relationship between waking preoccupations and dream content.

When it comes to people under the age of 15 or 16, we have little confidence in our norms in the categories of social interactions and misfortunes. We believe Foulkes (1982) has demonstrated dramatic developmental changes in dream content between the ages of 3 and 15, and that the dream life of children has less aggression and misfortune than Hall's sample led us to believe. Foulkes's findings support some of the gender differences and developmental trends in the Hall data, but they also support Foulkes's claims that the Hall sample is selective for dramatic intensity. Hall (1953c) and Hall and Domhoff (1963b) are guilty of exaggeration.

We conclude that children are important subjects for studying the cognitive development of dreaming as a psychological process, but not good subjects for studies of meaning in dream content. Foulkes's finding of poor recall with young children, along with the difficulties of obtaining permission to use children as subjects and then gaining their cooperation, convince us that the primary focus of content analysts should be on those people whose dreaming capabilities are fully developed and who are interested in participating in dream content studies. Still, it would be useful to know if the Most Recent Dream methodology could yield consistent findings with older children and teenagers.

We thus repeat what we think is the most important conclusion of this chapter: the findings on the dream reports of adult subjects show that our college student norms can serve as a basis for studies of dream meaning with unique adult groups or individual dream series. Before we turn to examples of such studies, however, we need to see what we learn by comparing our norms with findings from a wide range of nations and small preliterate cultures.

Continue on to Chapter 6

Go back to the Dream Library index.

dreamresearch.net home page dreamresearch.net contact info